paper-with-me

홈 › Papers

Mix-up Self-Supervised Learning for Contrast-agnostic Applications

2022-04-02 · Yichen Zhang, Yifang Yin, Ying Zhang, Roger Zimmermann

Contrastive self-supervised learning has attracted significant research attention recently. It learns effective visual representations from unlabeled data by embedding augmented views of the same image close to each other while pushing away embeddings of different images. Despite its great success on ImageNet classification, COCO object detection, etc., its performance degrades on contrast-agnostic applications, e.g., medical image classification, where all images are visually similar to each other. This creates difficulties in optimizing the embedding space as the distance between images is rather small. To solve this issue, we present the first mix-up self-supervised learning framework for contrast-agnostic applications. We address the low variance across images based on cross-domain mix-up and build the pretext task based on two synergistic objectives: image reconstruction and transparency prediction. Experimental results on two benchmark datasets validate the effectiveness of our method, where an improvement of 2.5% ~ 7.4% in top-1 accuracy was obtained compared to existing self-supervised learning methods.

📄 PDF Abstract BibTeX arXiv:2204.00901

Code (0)

등록된 구현이 없습니다.

Tasks

image-classificationImage ClassificationImage ReconstructionMedical Image Classificationobject-detectionObject DetectionSelf-Supervised Learning

Similar Papers 제목 키워드 기반

Contrastive Neural Processes for Self-Supervised Learning

2021-10-24 · Konstantinos Kallidromitis, Denis Gudovskiy, Kazuki Kozuka, Iku Ohama 외

Recent contrastive methods show significant improvement in self-supervised learning in several domains. In particular, contrastive methods are most effective where data augmentation can be easily constructed e.g. in comp…

Contrastive LearningData AugmentationSelf-Supervised LearningTime Series+2

CCGL: Contrastive Cascade Graph Learning

2021-07-27 · Xovee Xu, Fan Zhou, Kunpeng Zhang, Siyuan Liu

Supervised learning, while prevalent for information cascade modeling, often requires abundant labeled data in training, and the trained model is not easy to generalize across tasks and datasets. It often learns task-spe…

Data AugmentationGraph LearningGraph Representation LearningInformation Cascade Popularity Prediction+2

Foundation Models for ECG: Leveraging Hybrid Self-Supervised Learning for Advanced Cardiac Diagnostics

2024-06-26 · Junho Song, Jong-Hwan Jang, Byeong Tak Lee, DongGyun Hong 외

Using foundation models enhanced by self-supervised learning (SSL) methods presents an innovative approach to electrocardiogram (ECG) analysis, which is crucial for cardiac health monitoring and diagnosis. This study com…

Contrastive LearningDiagnosticSelf-Supervised Learning

Score-based Self-supervised MRI Denoising

2025-05-08 · Jiachen Tu, Yaokun Shi, Fan Lam

Magnetic resonance imaging (MRI) is a powerful noninvasive diagnostic imaging tool that provides unparalleled soft tissue contrast and anatomical detail. Noise contamination, especially in accelerated and/or low-field ac…

DenoisingDiagnosticSelf-Supervised Learning

Towards Domain-Agnostic Contrastive Learning

2020-11-09 · Vikas Verma, Minh-Thang Luong, Kenji Kawaguchi, Hieu Pham 외

Despite recent success, most contrastive self-supervised learning methods are domain-specific, relying heavily on data augmentation techniques that require knowledge about a particular domain, such as image cropping and …

Contrastive LearningData AugmentationImage CroppingRepresentation Learning+1